NDSS

DIAT: Data Integrity Attestation for Resilient Collaboration of Autonomous Systems

Tigist Abera (Technische Universität Darmstadt), Raad Bahmani (Technische Universität Darmstadt), Ferdinand Brasser (Technische Universität Darmstadt), Ahmad Ibrahim (Technische Universität Darmstadt), Ahmad-Reza Sadeghi (Technische Universität Darmstadt), Matthias Schunter (Intel Labs)

Networks of autonomous collaborative embedded systems are emerging in many application domains such as vehicular ad-hoc networks, robotic factory workers, search/rescue robots, delivery and search drones. To perform their collaborative tasks the involved devices exchange various types of information such as sensor data, status information, and commands. For the correct operation of these complex systems each device must be able to verify that the data coming from other devices is correct and has not been maliciously altered.

In this paper, we present DIAT – a novel approach that allows to verify the correctness of data by attesting the correct generation as well as processing of data using control-flow attestation. DIAT enables devices in autonomous collaborative networks to securely and efficiently interact, relying on a minimal TCB. It ensures that the data sent from one device to another device is not maliciously changed, neither during transport nor during generation or processing on the originating device. Data exchanged between devices in the network is therefore authenticated along with a proof of integrity of all software involved in its generation and processing. To enable this, the embedded devices’ software is decomposed into simple interacting modules reducing the amount and complexity of software that needs to be attested, i.e., only those modules that process the data are relevant. As proof-of-concept we implemented and evaluated our scheme DIAT on a state-of-the-art flight controller for drones. Furthermore, we evaluated our scheme in a simulation environment to demonstrate its scalability for large-scale systems.